The SARS-CoV-2 virus, responsible for the COVID-19 pandemic, has continuously mutated, resulting in numerous variants. The spike protein (S), a crucial component mediating viral entry into human cells, is the primary target for neutralizing antibodies. This protein, a homotrimer with S1 and S2 subunits, undergoes extensive N-glycosylation at 22 potential sites per protomer. N-glycosylation plays vital roles in protein folding, stability, and immune evasion by creating a "glycan shield". Variants of concern (VOCs), including Alpha, Beta, Gamma, and Delta, display enhanced infectivity and immune evasion, likely due to mutations in the spike glycoprotein. These mutations might alter the glycosylation profile of the spike protein, affecting its function and the efficacy of vaccines and therapeutics. This study focuses on investigating whether these VOCs have altered glycosylation profiles compared to the D614G variant, providing valuable data for understanding SARS-CoV-2 and developing effective countermeasures.

Extensive research has utilized advanced mass spectrometry (MS) techniques to analyze N- and O-glycoforms of SARS-CoV-2 S proteins. Studies have predominantly used recombinant protein constructs expressed in various cell types (human, monkey, insect). While all 22 N-glycosylation sequons have been shown to be occupied, O-glycosylation occupancy is relatively low. The S ectodomain is largely modified by complex carbohydrates, with some oligomannose structures present. Previous research has highlighted the roles of N-glycans in maintaining protein structure, shielding epitopes, and potentially influencing RBD conformation. Studies have also demonstrated that glycosylation deletions or interference with host N-glycosylation reduce viral production and infection spread. However, inconsistencies exist in glycosylation analysis results across different laboratories, likely due to variations in protein constructs, sample preparation, and data analysis methods. This study addresses these inconsistencies by employing a standardized approach and multiple digestion strategies.

Recombinant poly-histidine-tagged, trimeric ectodomains of SARS-CoV-2 S variants (D614G, Alpha, Beta, Gamma, and Delta) expressed in HEK293 cells were purchased from a single manufacturer (R&D Systems) to ensure consistency. Multiple protease digestion strategies (single and dual enzyme digestions using trypsin, Lys-C, chymotrypsin, alpha-lytic protease, and Asp-N) were employed to generate glycopeptides with single sequons for accurate quantification. Digestion involved denaturation, reduction, alkylation, and enzymatic digestion at varying temperatures and times. Nanoflow liquid chromatography coupled to electrospray ionization tandem-mass spectrometry (LC-MS/MS) analysis was performed on an Orbitrap Eclipse Tribrid mass spectrometer. A signature ion triggered EThcD method was used, with HCD and EThcD fragmentation. Data was processed using PMi-Byonic within Proteome Discoverer. Byonic searches were conducted with specific cleavage sites for each enzyme, allowing up to three missed cleavages and accounting for variable modifications (deamidation, oxidation). A glycan database containing human N-glycans was used. Relative glycan abundance at each site was calculated based on normalized peak intensity ratios, and data from multiple digestions were rigorously validated and selected using defined criteria.

N-glycans were detected at all 22 conventional glycosylation sites, with some exceptions due to mutations (N17 in Delta, unresolved N17 in Gamma, insufficient data for N149 and N717). Two novel glycosylation sites (N20 and N188) in Gamma were identified. The study used multiple protease digestion methods to obtain robust data; no single condition worked for all sites. Complex N-glycans predominantly occupied most sites in all variants, except N61 and N234 which had high oligomannose. High similarity in N-glycan processing states (paucimannose, oligomannose, hybrid, complex) was observed between variants. Delta showed lower complex glycan abundance at many sites compared to others. The study compared its results to previously published work from Kuo et al., Newby et al., and Shajahan et al., noting both consistencies (predominance of complex glycans) and discrepancies (variations in glycan types at specific sites). The analysis focused on microheterogeneity, detailing the relative abundance of top N-glycans at each site, revealing variations in distribution between variants, particularly at N122, N165, N717, N1158, and N1173. The number of distinct N-glycans per site varied from 24 to 91. Fucosylation and sialylation levels varied across sites, with some showing high fucosylation (N343, N603) and low sialylation, potentially influencing immune evasion. The two novel Gamma glycosylation sites (N20 and N188) were found to be almost fully occupied, with N20 primarily having complex glycans and N188 dominated by oligomannose glycans.

This study's findings contribute significantly to our understanding of SARS-CoV-2 spike protein glycosylation and its evolution. The high similarity in N-glycosylation profiles across variants suggests that this modification is crucial for spike protein structure and function, maintaining its stability and immune-shielding properties. Variations observed in certain NTD sites might impact interactions with ACE2, warranting further investigation. The results also support the continued effectiveness of existing vaccines, although subtle differences in glycan composition could affect efficacy. The multiple digestion approach helped to overcome inconsistencies in previous studies, providing a more comprehensive and robust analysis. The identification of the novel glycosylation sites in the Gamma variant offers further insight into the virus's adaptation mechanisms.

This study provides a comprehensive, high-resolution analysis of N-glycosylation in four SARS-CoV-2 VOC spike proteins. The high degree of similarity in glycosylation across variants highlights the importance of this modification for viral function and the potential continued effectiveness of current vaccines. Future research could focus on exploring the functional implications of observed glycan variations, particularly those impacting receptor binding and immune evasion. Further investigation into O-glycosylation, other post-translational modifications, and the impact of different expression systems would also enhance understanding.

The study's limitations include the potential omission of glycosylation at some sites due to the selection criteria used, despite multiple digestion strategies. The analysis did not characterize other glycan types (e.g., O-glycosylation, sulfated, phosphorylated glycans). The lack of determination of specific structural features of glycoforms (e.g., glycan linkages) and potential biases due to recombinant protein source, MS acquisition parameters, and data processing algorithms are also noted. The use of a single batch of recombinant proteins from one manufacturer might also introduce biases.